code of learn deep learning with pytorch

PyTorch

Learn Deep Learning with PyTorch

Thank you very much for purchasing this book. This github repository contains the example code for PyTorch, which is an introduction to deep learning. Due to my limited level, I referred to some online information when writing this book, and I would like to express my respect to them here. As deep learning technology is developing rapidly, PyTorch is also constantly updated, and I have many fields that I have not covered when I completed this book, so this repository will be updated as a follow-up service for purchasing the second book. I hope I can provide you with slim power on the road to getting started with deep learning.

Note: Due to the change of PyTorch version, the code in the book may have bugs, so all codes are mainly based on the ones in this github.

Configure the environment

The book has already given in detail how to configure the python environment based on Anaconda and install PyTorch. If you use your own computer and have an Nvidia graphics card, then you can happily enter the world of deep learning. If you don’t have an Nvidia graphics card, then we need a cloud computing platform to help us learn the deep learning journey. How to configure the AWS computing platform

The following course catalogs are different from those in the book, because the content is being updated to the second edition, and the second edition will be launched soon! !

Course Catalog

part1: Basics of Deep Learning

• Chapter 2: PyTorch Basics

  • Tensor and Variable
  • Automatic guidance mechanism
  • Dynamic and static images

• Chapter 3: Neural Networks

  • Linear model and gradient descent
  • Logistic regression and optimizer
  • Multi-layer neural network, Sequential and Module
  • Deep neural network
  • Parameter initialization method
  • Optimization algorithm
    • SGD
    • Momentum method
    • Adagrad
    • RMSProp
    • Adadelta
    • Adam

• Chapter 4: Convolutional Neural Networks

  • Convolutional modules in PyTorch
  • batch normalization
  • Deep network using repeating elements, VGG
  • More rich network structure, GoogleLeNet
  • Deep residual network, ResNet
  • DenseNet
  • Better training convolutional networks
    • Data Enhancement
    • Regularization
    • Learning rate decay

• Chapter 5: Recurrent Neural Network

  • Recurrent neural network modules: LSTM and GRU
  • Using RNN for image classification
  • Time series analysis using RNN
  • Applications of natural language processing:
    • Word Embedding
    • N-Gram Model
    • Seq-LSTM to make part-of-word prediction

• Chapter 6: Generate adversarial networks

  • Automatic encoder
  • Variational automatic encoder
  • Generate adversarial networks
  • Deep Convolutional Adversarial Networks (DCGANs) Generate Human Faces

• Chapter 7: Deep reinforcement learning

  • Q Learning
  • Open AI gym
  • Deep Q-networks

• Chapter 8: PyTorch Advanced

  • tensorboard visualization
  • Flexible data reading introduction
  • Introduction to autograd.function
  • Data parallelism and multi-GPU
  • Convert to Caffe2 model using ONNX
  • How to deploy a trained neural network
  • Create your own PyTorch usage habits

part2: Application of deep learning

• Chapter 9: Computer Vision

  • Fine-tuning: Transfer learning through fine tuning
  • First experience of kaggle: Cat and dog battle
  • Semantic segmentation: Implement pixel-level classification through FCN
  • Pixel to Pixel Generative Adversarial Networks
  • Neural Transfer: Implementing Style Transfer through Convolutional Networks
  • Deep Dream: Exploring the world in the eyes of convolutional networks

• Chapter 10: Natural Language Processing

  • Char RNN implements text generation
  • Image Caption: Implement image subtitle generation
  • seq2seq implements machine translation
  • cnn + rnn + attention to realize text recognition

Some other resources

For some public courses and learning resources for deep learning, please refer to my repository

You can follow my Zhihu column and blog, and you will often share some deep learning articles

Resources about PyTorch

My github repo pytorch-beginner

pytorch-tutorial

the-incredible-pytorch

practical-pytorch

PyTorchZeroToAll

Awesome-pytorch-list

Acknowledgement

Some parts of the second edition of this book refer to the Chinese tutorial of mxnet gluon, and use MXNet/Gluon to learn deep learning.

Gluon is a framework that is very similar to PyTorch. It is very simple and easy to use. I recommend you to learn it and also recommend gluon's Chinese courses, which are taught in all Chinese, have videos, and code practices. It can be said to be the most comprehensive Chinese deep learning tutorial.